Flagella of a plant-growth-stimulating Pseudomonas fluorescens strain are required for colonization of potato roots

(1)

0021-9193/87/062769-05$02.00/0

Flagella of a

Plant-Growth-Stimulating

Pseudomonas fluorescens

Strain Are Required for Colonization of Potato Roots

LETTY A. DE WEGER,1* CORNELIA I. M. VAN DERVLUGT,1 ANDRE H. M. WIJFJES,1 PETER A.H. M.BAKKER,2 BOB SCHIPPERS,2 ANDBEN LUGTENBERG1

Departmentof PlantMolecularBiology, Botanical Laboratory, Leiden University, 2311 VJLeiden,1 and Willie Commelin

Scholten,

Phytopathological

Laboratory,

3742 CP

Baarn,2

TheNetherlands Received 9 January 1987/Accepted 18 March 1987

The role of

motility

in the colonization of potato roots by Pseudomonas bacteria was studied. Four TnS-induced flagella-less mutants of the plant-growth-stimulating P.fluorescens WCS374 appeared to be impairedintheir

ability

tocolonizegrowingpotato roots.

Certain fluorescent Pseudomonas isolates are able to increase plant growth (3, 10, 16, 28) or to protect plants against microbial pathogens (17,25).Our research is focused onstrains thatcanbe used toeliminate potatoyield reduc-tions as observed in Dutch fields in which potatoes are frequently grown (28). These yield reductions can be abol-ished by treatment of the seed potatoes with particular Pseudomonas strains (3, 16, 27, 28,). Essential for this beneficial effect of the bacteriaonplantsistheproduction of fluorescent siderophores,

Fe3+-chelating

compounds (3, 22) which supposedlyenable thePseudomonascells toscavenge most of the

Fe3`

ions from the Fe3+-poor soil, thereby depriving deleterious microorganisms of this essential ele-ment. A second factor whichissupposed to beessential for efficient protection of therootsagainstdeleterious microor-ganisms is the delivery of siderophoresalong thewholeroot systemof theplant, whichrequiresefficient colonization of the potatoroots. Althoughroot colonization isvery impor-tant in nature, virtually nothing is known about it at the molecular level. We studied the role of motility of Pseu-domonas bacteria in the colonization ofpotato roots.

MATERIALS ANIDMETHODS

Strains and growth conditions. The potato root isolate Pseudomonas fluorescens WCS374 (7, 9, 10) is resistantto nalidixic acid (25 ,ug/ml). Strain JM3741 is a TnS-marked

derivative of WCS374 which wasisolatedby J. Marugg. It does not differ significantly from the parental strain in its root-colonizing ability, its siderophore production, its

growthrate in the complex King B medium or inminimal medium, and its motility(P.A. H. M. Bakker,unpublished

data). Escherichia coli CSH52 harboring the mobilizable

plasmid pSUP202 (Apr Cmr Kmr Tcr) and strain S17-1 harboring pSUP2021 (=pSUP202 withTn5inserted into the gene for tetracycline resistance) were obtained from R. Simon(29). Pseudomonas strainswere grownat28°CandE.

ccli

strainsweregrown at37°C inKingBmedium (15) under

vigorous aeration. For the root colonization assay TnS-labeled strains were cultivated at28°C for 48 h on KingB medium solidified with 1.6% agar and supplemented with

kanamycin (50 ,ug/ml), TheTnS-labeled strains were

resist-anttokanamycin (200 ,ug/ml) and streptomycin (200 ,ug/ml)

*Correspondingauthor.

(TnS encodes both kanamycin and streptomycin resistance in Pseudomonas species).

Isolation ofTn5-induced nonmotile mutants. TnS-induced mutants wereobtained by the method of Simon et al. (29) with slightmodifications asdescribed by Maruggetal. (22). Briefly,E.coli S17-1harboring pSUP2021wasmated withP.

fluorescens WCS374 for 3 to 4 h at

28°C.

Transconjugants

were selected on King B agar plates containing 25 ,ug of

kanamycin per ml and 20 ,g of nalidixic acid per ml. Colonies were screenedformotility onmotility plates

con-sistingof 20-fold-dilutedKingBmediumsolidified with0.3%

agar. Afterspotinoculation, motilitywasjudged after incu-bation for16 h at28°C.

Isolation of flagella. Themethod described byDePamphilis andAdler (6) forisolation of flagellawasslightly modified.

Late-logarithmic-phasecells wereharvestedand gently sus-pended in 0.1 MTrishydrochloride (pH 7.8)at adensity of

1010

cells per ml. Deflagellation with a Sorvall Omnimixer

with the speed control dial at position4 for 15 min at 0C

resulted in loss of motilityfor99% of the cells, without loss ofviability. The suspension was centrifuged at 12,000 x g for 10 min, and the resulting supernatant fluid was centri-fugedat90,000x gfor2htoharvesttheflagella. Flagellato beusedfor immunization ofarabbitwerefurtherpurifiedby isopycnic density centrifugation in cesium chloride. The

flagella were suspended in 5 ml of 50% (wt/vol) cesium chloride andlayeredbetween 5mlof40% and5ml of60% cesium chloride. Aftercentrifugationat 100,000 x gfor3 h atroomtemperaturethe gradientwasfractionated,andthe flagellar layer was identified by its turbidity. For density

measurements 0.2-mlsamples were weighed. Cesium chlo-ridewasremovedfrom theflagellarfractionbydialysis.

Preparation of flagellum-specific antiserum. A rabbit was

immunizedintradermally with100 ,ugof thepurifiedflagellar

preparation emulsified in Freund complete adjuvant. Three boosterinjections inFreundincomplete adjuvantweregiven

at 2-weekintervals. Serum was obtained2 weeks after the lastinjection and stored at -20°C. Theresulting antiserum

cont4ined

antibodies reacting (in immunoblots) with the

flagellar subunit and with lipopolysaccharide (LPS). To obtain a flagellum-specific antiserum, the LPS antibodies were removed byabsorbing the antiserumwith cells of the nonmotile mutant LWM74-29. Approximately

1010

cells washed with 10 mM phosphate-buffered saline (pH 7.2) containing1mgofNaN3per mlweresuspendedin0.5mlof antiserum. After incubation for 1 h at

37°C,

cells were 2769

on January 16, 2017 by WALAEUS LIBRARY/BIN 299

http://jb.asm.org/

(2)

removed by centrifugation. The procedure was repeated with fresh cells. After this treatment the reaction with LPS onimmunoblotswasnolongerobserved, whilethereactivity toward the flagellar subunit had notbeen affected.

SDS-PAGE.Samplesweresolubilizedby incubation for15 min at95°C inthestandard sample mixture and subjectedto sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as described previously (18). Cell envelope samples were obtained by differential centrifugation after disruption of the cells by ultrasonictreatment(18). Samples containing10 to 20,ugproteinwereappliedperlane, and gels were stained with fastgreen FCF (18). Forthe analysis of LPS the samples were first treated with proteinase K (13)

after which 10-fold dilutions were applied per lane. Gels were stained with silver reagent (31).

Immunoblotting. Forimmunoblottingthe amountsapplied perlanewere 100-fold lessthan theamountsappliedtogels

to be stained with fast green. For analysis of the culture supernatant, 2,ulfromastationary-phaseculture(A620=5 to 6) was applied perlane. Electrophoretically separated pro-teins were electrophoretically transferred to nitrocellulose with a blotting cell (Bio-Rad Laboratories, Richmond, Calif.).Thenitrocellulose sheetwasthenincubated with the flagellum-specific antiserum, followed by incubation with goat anti-rabbit antiserum-alkaline phosphatase conjugate.

Theblotting reactionwasvisualized with fast redTRsalt and naphthol AS-MX phosphate (Sigma Chemical Co., St. Louis, Mo.) asdescribed by vanderMeide etal. (33).

DNA isolation and analysis. DNA of small plasmids was

isolated bythemethod of Birnboim andDoly(4)andpurified by isopycnic density centrifugation in cesium chloride. Ex-traction of total DNA and digestions with restriction endonucleases (BamHlandEcoRI)were doneasdescribed by Maniatis et al. (20). DNA fragments were transferred from the 0.8% agarose gels to nitrocellulose filters (20).

32P-labeled

plasmidpNP520 (23) was used as aTnS probe, and pSUP202 (29) was used as a probe for vector plasmid

DNA. Conditions for hybridization with 32P-labeled DNA probes prepared by nick translation were as described by

Maniatisetal. (20).

Electron microscopy. Bacterial cells were negatively stained with phosphotungstic acid and examined with a

Philips 300transmission electronmicroscope.

Root colonization assay. Potato stem cuttings with roots

approximately 1 cm long were dipped in bacterial suspen-sions of108 CFU/ml. After removal ofexcesssuspensionby shaking, the stem cuttings were embedded in clay soil in tubes ofpolyvinyl chloride (4.5-cm diameter; 12 cm long). Thissoil wascollected fromafield inwhichpotatoes were grown once every 6 years. The soil moisture content was

brought to 25% about 24 h before the soil was transferred into the polyvinyl chloride tubes. Before the stem cuttings were planted, the tubes were placed on a layer of wet vermiculite (Fig. 1), and this system was allowed to equili-brate for24h. During the growth oftheplantsthehumidity ofthe soil wasregulated by this layerofvermiculite, so a directmovementofwaterfromtop to bottom was avoided. After12days of growthin agreenhouse,root samples 1 cm long (totalroot fresh weight, 0.3 g per sample) were taken fromthree differentdepths (0to1,4, and 8 cm) and shaken in 5 ml of 0.1% Proteose Peptone (Difco Laboratories, Detroit, Mich.) with2.5 gofglass beads (3-mm diameter) in aVortexmixeratmaximumspeed. Appropriatedilutionsof thesesuspensionswereplatedonKingBagarplates supple-mented with 200 ,ug of kanamycin per ml and 200 ,ug of

streptomycin perml. Thenumber of colonies was counted

after48 hofgrowth.Results wereanalyzed bythe Kruskal-Wallistestfollowedby nonparametricmultiple comparisons bythe simultaneoustestprocedure (30).

Protein determination. Protein was determined by the methoddescribedbyMarkwelletal.(21)with bovineserum

albuminasthe standard.

RESULTS

Isolation of nonmotile TnSmutants.Matingof E.coliS17-1

containing the plasmid

pSUP2021

with P. fluorescens

WCS374 resulted in kanamycin-and nalidixic acid-resistant

(KanrNalr) transconjugants at afrequency of 2 x 10-6per

recipientcell, whereasthefrequencywasless than 10-9 for spontaneousmutants. KanrNalrmutants of strainWCS374 were screened formotility on semisolid medium. The wild-type strain WCS374 formed swarms that were 2 cm in

diameter, while the growth areaof nonmotile mutants was

restricted to the spot of inoculation. Of 500 Kanr Nalr mutants tested, 4 nonmotile mutants were obtained,

desig-nated as strains LWM74-4, LWM74-29, LWM74-30, and LWM74-36.

These four mutants were also nonmotile when observed

by phase-contrastmicroscopy. Using electronmicroscopy, up to nine polar flagella were observed for the wild-type

strain WCS374, while no flagella were observed for the nonmotilemutantsLWM74-4, LWM74-29, and LWM74-30. Mutant strain LWM74-36 occasionally contained one fla-gellumpercell(datanotshown).

The nonmotile mutant strains were not affected in (i) siderophore production, (ii) growthrate in minimalmedium

orthecomplexKingBmedium, and(iii) membraneprotein patternsorLPS ladderpatternsasanalyzed bySDS-PAGE

(datanotshown).

Isolation and characterization of flagella. After removal from the cells by shearing, flagella wereisolated by differ-ential centrifugation. SDS-PAGE showed thatthe resulting preparation contained a dominant 58,000-dalton protein as

wellas alargenumberof minorprotein bands(Fig. 2, lane

0-1 cm

-4 cm -8 cm

wet-- _ g

vermiculite

FIG. 1. Schematicrepresentationof thebioassay usedto deter-minerootcolonization. The potatostemcuttingswereinoculatedas described in the text, after whichtheywereallowedtogrowina polyvinylchloride(PVC)tubecontainingsoil whichwasplacedon top ofalayerofwetvermiculite. After 12 days of growth, root sectionsat0-to1-,4- and8-cmdepthsweretakentodetermine the numberofviable bacteria presentonthedifferent parts of theroot.

http://jb.asm.org/

(3)

1). Subsequent isopycnic density gradient centrifugation in cesium chloride of this flagellarpreparation resultedinloss of minorprotein bands with higher molecular sizes than the 58,000-dalton protein, but in the middle and lower part of the gel minor protein bands were still visible (data not shown). The fraction containing the 58,000-dalton protein layered at adensityof 1.32 to 1.34g/ml,which is close to the density of flagella of Bacillus subtilis and E. coli (1.30 g/ml [6]). Therefore, the 58,000-dalton protein most probably repre-sents the subunit of the flagella of strain WCS374.

Cells of the fournonmotile mutants were sheared, and the suspensions were differentially centrifuged. The resulting pelletscontained two to four times less protein per volume of cell culture compared with the flagellar pellet of the wild-type strain. The resulting preparations were adjusted to the same protein concentration and analyzed by SDS-PAGE. These preparations of the nonmotile mutantsdidnot reveal the heavy 58,000-dalton protein band which was observed in similar preparations of the wild-type strain, but the minor proteinbands were present (Fig. 2, lanes 2 to 5).

Molecular characterization of mutations. Total DNA iso-lated from the four nonmotile mutants was digested with restriction endonucleasesEcoRI(for whichTnSDNA has no restriction site) and BamHI (which cuts TnS at a site 2.65 kilobases from one end and 2.95 kilobases from the other).

A B

58kD - _

C

--I

1 121

31

1

2131 1 112131 FIG. 3. Immunoblotsofresuspended flagellar pellets (A),whole cells (B), and culture supernatant fluids (C). Lanes 1, wild-type

strain WCS374; 2 and 3, nonmotile mutants LWM74-36 and LWM74-30, respectively.MutantstrainsLWM74-4and LWM74-29, hadbehavior identicaltothat of strainLWM74-30in immunoblots. Thepositionofflagellinis indicatedbyan arrow(58kD).

66K- 55K- 45K-OM mow -a *1 "A a .!Xm 36K- 29K- 24K-14K- .C

1

1213141,51

FIG. 2. SDS-PAGE of theflagellaof strain WCS374(lane 1)and of the comparable preparations of the nonmotile mutant strains LWM74-4 (lane 5), LWM74-29 (lane 4), LWM74-30 (lane 3), and LWM74-36 (lane 2). Equalamounts(5 ,ug)ofproteinwereappliedto

eachlane. Thepositionofflagellinis indicatedbyan arrow(58kD)

attheright, and thepositionsof molecularweightstandardproteins

are indicated at the left (k, 103). Note that the position of the 58,000-dalton (58kD) proteinband in theflagella preparationof the

wildtype(lane 1) isinbetweenthepositions oftwoproteinspresent in thepreparationsof the four nonmotilemutants(lanes2to5).

Digests were fractionated by electrophoresis, blotted to nitrocellulose, and hybridized with the 32P-labeled TnS probe.Theresultingautoradiogram showed thatin allEcoRI digests onlyonebandhybridized withTnS(datanotshown). Moreover, fractionated EcoRI and BamHI digests of the fourmutants showed differentbandsforeachofthe mutant strains. Hybridization with the 32P-labeled pSUP202 probe (=pSUP2021 withoutthe TnS) didnot reveal any bandson the autoradiogram. These results show that the four TnS mutationsareindependent, thatthey arelocatedindifferent restrictionnucleasefragments, andthat no vector DNA was transferredtothe mutants.

The flagellar preparation ofthe wild-type strain and the comparable preparations of the mutant strains were sub-jectedtoSDS-PAGE, andtheflagellinsubunitwasdetected by immunoblotting with a flagellum-specific antiserum. In theflagellar preparation of thewild-typestrain WCS374the

flagellarsubunit wasdetected. Inaddition, several proteins with lower molecular weights (Fig. 3A, lane 1) were de-tected. Inthe lanes with the "flagellar preparations" ofthe fournonmotilemutantsnoreactionwasobservedexcept for afaint band atthe gel front(Fig. 3A, lanes2 and 3).

Using theflagellum-specific antiserum inanimmunoblot,

flagellinwas notonly detectedinwhole cells of thewild-type strainWCS374,butalsointhoseofmutantstrain LWM74-36 (Fig. 3B).Noflagellinwasdetectedincells oftheotherthree nonmotile mutants. Flagellin appeared to be present in substantial amounts in the culture supematant of strain LWM74-36(Fig. 3C,lane2),whileonlyasmallamountwas detected in the culture supematant of the wild-type strain WCS374. Noflagellinwasfound intheculture supernatants

http://jb.asm.org/

(4)

TABLE 1. Rootcolonization ofpotatoplantsinoculated with P.fluorescensJM3741 and with nonmotilemutantsa

Log1oof no. of CFU at a depth (cm)of: Strain

Oto1 4 8

JM3741 7.4 ± 0.5(a) 5.0± 0.5(a) 4.5 ±0.4(a) LWM74-4 7.4±0.3(a) 4.7± 1.9(a) 1.7 ±2.2 (b) LWM74-29 7.2± 0.7(a) 4.3 ± 1.6(a) 1.2 ±2.0 (b) LWM74-30 7.2± 0.4 (a) 2.9±2.4(b) 1.6 ±2.0(b) LWM74-36 6.7± 0.2 (a) 1.6 ± 2.1(c) 0.8 ± 1.8(b)

aThenumbers ofkanamycin-andstreptomycin-resistantCFU on the roots atvarious depths (seeFig. 1) weredetermined. The values given are mean values±standarddeviations of log-transformeddeterminationsof 10

repli-cates. Within a givencolumn values with the same letter inparenthesesare

notsignificantly different at P = 0.05, based on nonparametric multiple

comparisonby the simultaneous testprocedure.

of the other three nonmotile mutant strains (Fig. 3C). Be-sides the 58,000-dalton flagellin band, another protein of approximately 55,000 daltons reacted with the flagellum-specific antiserum in the lane containing the culture super-natant fluid of strain LWM74-36 (Fig. 3C, lane 2).

SDS-polyacrylamide gel analysis showed that the amounts of these two proteins varied with the age of the cultures. In cultures with an optical density at 620 nm below 4, the

55,000-daltonproteinshowed thehighest intensity, while in cultureswithanoptical density at620nmhigherthan 5the

58,000-daltonprotein bandwasdominant(datanot shown). Theflagellin in the culture supernatant of the wild-type strain

appeared tobe sedimentable byultracentrifugation (90,000

xg for 2 h) asexpectedforflagellar fragments, whereasthe

flagellinofmutantstrain LWM74-36 remainedpresentin the

supernatant fluid under these conditions (data not shown), indicatingthatitwaspresent asnonpolymerized subunits.

Colonizationofpotatoroots. Theabilities of the

wild-type

strain and the nonmotile mutants to colonize potato roots werecompared withpotato stemcuttings inoculated with the

motile TnS-labeled strain JM3741 or with one of the nonmotile mutants. After 12 days of growth in soil, the number ofTn5-carrying CFU on various parts of the root system (Fig. 1) was determined (Table 1). The numbers of

TnS-containing CFU on root samples tak-en from the 0 to 1-cm

depth

didnotdiffersignificantlyamongtheplants. This wastobeexpected sinceinoculation took place atthispart of therootsystem.However, in rootsamples taken froma part of the root system formed after inoculation (e.g., root samples taken from the 8-cm depth), the number of

TnS-containingCFU wasdrastically reduced in the plants inoc-ulatedwith nonmotile mutants whencomparedwith that in plants inoculated with the wild-type strain (Table 1). For mutant strains LWM74-30 and LWM74-36 the number of

TnS-containingCFU wasalso significantlyreduced on root samples taken from the 4-cm depth. The results show that

motilityisrequiredtocolonizegrowingroots successfully. DISCUSSION

Flagellin

of P.

fluorescens WCS374. The

plant-growth-stimulating strain WCS374 isamotileP.fluorescens strain withup to nine polarflagellapercell(datanot shown). Of 500TnS-inducedmutantsof this strain,4nonmotilemutants were isolated.A largenumberofgenes are involvedin the formation of functional flagella of gram-negative bacteria such as E. coli(2), Salmonella typhimurium(24), and Pseu-domonas aeruginosa (32). The high frequency with which the nonmotile mutants of this P.

fluorescens

strain were

foundaswellastheobservationthattheinsertions ofTn5in

the four nonmotile mutant strains are located on four dif-ferent DNA restriction fragments are consistent with the large number ofgenesrequiredfor thesynthesisand assem-bly of flagella.

Theflagellar preparation of thewild-type strainshoweda

dominant 58,000-dalton protein band and minor protein

bands with lower molecular sizes(Fig. 2).This58,000-dalton

proteinmostlikely represents the flagellar subunitflagellin

since(i)itcopurifieswithparticles withabuoyantdensity of

1.32 to 1.34 g/ml, which is closeto thebuoyantdensity of flagella ofB. subtilisandE. coli(1.30g/ml[6]); (ii)itisthe

major protein in purified flagella; and (iii) it is absent in

comparablepreparationsof thenonflagellatedmutants(Fig.

2 and 3). The identity of the other protein bands in the

flagellar preparations(Fig. 2) isnotknown. At leastsomeof thesemostlikelyrepresentdegradation productsofflagellin sincetheyweredetectedbyimmunoblottingwith

flagellum-specificantiserum in theflagellar

preparation

of the wildtype

(Fig.

3).

Characterization of defectsofnonmotilemutants. Inoneof the fournonmotile mutants(LWM74-36), the58,000-dalton flagellin protein was present in intact cells andas

nonpoly-merizedsubunitsinthe cell culture supernatant fluid (Fig.3).

In this mutant strain, the impairment in biosynthesis of

flagellamaybe in thebasalbodyorhookstructure,thusnot

supplying

afunctional basis for the polymerization of the

flagellar subunits into a flagellum, comparable to the S. typhimurium mutantsdefective in hook-associatedproteins

(14). Alternatively, LWM74-36 may be defective in the

polymerization processitself. The flagellin was usually

se-creted into the growth medium, but occasionally mutant

LWM74-36cellssynthesized single polar flagella as seenby

electronmicroscopical examination (datanotshown). Examination of the nonmotile mutants LWM74-4,

LWM74-29, and LWM74-30 by immunoblotting showed no

detectable

flagellin

incompletecellsorin the culture super-natant of these strains.

Therefore,

these three nonmotile

mutants

appe?ar

impaired in the synthesis of

wild-type

flagellin.

Roleof

motility

inrootcolonization. Previous studies had shownthat there is chemotactic attraction ofrhizobacteria

byrootexudate(5, 12),rootmucilage

(19),

and seedexudate (26). Furthermore, P. aeruginosa cells movethroughmoist soil adistance of2cmin24h(11). Scheretal. (26) showed

that Pseudomonas putida could move 2 cm in 48 h to

imbibing soybean seeds in raw soil. These studies indicate that chemotactic movement ofbacteria in the rhizosphere

cantakeplaceandsuggestthatmotilitycanplayarole in the

colonizationofroots.

Theroot-colonizingabilities of thenonmotilemutantsand the

wild-type

strainWCS374werecomparedinagreenhouse

root colonization assay. In this assay the soil was not

irrigatedfromthetop,but soil humiditywasprovidedfrom below (Fig. 1). Each ofthe four nonmotile mutants had a

reduced abilitytocolonize growingroots(Table 1). Before

drawingtheconclusion that

motility

playsanessential role in efficientrootcolonization,wehavetoconsider the

possibil-ity that the mutations leading to the

nonmotile

phenotype have pleiotropic effects. The

chvB

mutation in Agrobac-terium tumefaciens, which causes defective attachmentto

plant tissues, alsoappears toaffect the formationofflagella (8).

Similarly,

thenonmotilephenotype inE. coliwasfound amongmutantslackingtheLPSsugarheptose (1).However, no evidence has been found for pleiotropic effects in our mutants to date since they have

wild-type

properties with

http://jb.asm.org/

(5)

respect to composition ofthe LPSs and proteins in the cell envelope, production of siderophores, and growth rate in King B medium or minimal medium. Moreover, the Tn5 insertions in the four nonmotile mutants are independent, and it isnot likely that the lack of motility in each of these mutants is due to different pleiotropic mutations which by themselves affect the root-colonizing ability of the strain. The results of this study therefore strongly suggest that

flagella ofP.fluorescens WCS374 play an essential role in the colonization ofpotato roots.

ACKNOWLEDGMENTS We thank Carel Wijffelman for helpful discussions.

Thisinvestigation was supported by the Netherlands Technology Foundation (STW).

LITERATURE CITED

1. Ames, G. F., E. N. Spudich, and H. Nikaido. 1974. Protein composition of the outer membrane of Salmonellatyphimurium: effect of lipopolysaccharide mutations. J. Bacteriol. 117:406-416.

2. Bachmann, B. J., and K. B. Low. 1980. Linkage map of Escherichiacoli K-12, edition 6. Microbiol. Rev. 44:1-56. 3. Bakker, P. A. H. M., J. G. Lamers, A. W. Bakker, J. D.

Marugg, P. J. Weisbeek, and B. Schippers. 1986. The role of siderophores in potato tuber yield increase by Pseudomonas putida in a short rotation of potato. Neth. J. Plant. Pathol. 92:249-256.

4. Birnboim, H. C., and J. Doly. 1979. A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Res. 7:1513-1523.

5. Currier, W. C., and G. A. Strobel. 1977. Chemotaxis of Rhizo-bium spp. toaglycoprotein produced by birdsfoot trefoil roots. Science 196:434-436.

6. DePamphilis, M. L., and J. Adler. 1971. Purification of intact flagella from Escherichiacoli and Bacillus subtilis. J. Bacteriol. 105:376-383.

7. De Weger, L. A., R. van Boxtel, B. van der Burg, R. Gruters, F. P. Geels, B. Schippers, and B. Lugtenberg. 1986. Siderophores and outer membrane proteins of antagonistic plant-growth-stimulating, root-colonizing Pseudomonas spp. J. Bacteriol. 165:585-594.

8. Douglas, C. J., R. J. Staneloni, R. A. Rubin, and E. W. Nester. 1985. Identification and genetic analysis of an Agrobacterium tumefaciens chromosomal virulence region. J. Bacteriol. 161:850-860.

9. Geels, F. P., and B. Schippers. 1983. Selection of antagonistic fluorescentPseudomonas spp. and their root colonization and persistence following treatment of seed potatoes. Phytopathol. Z. 108:193-206.

10. Geels, F. P., and B. Schippers. 1983. Reduction of yield depres-sions in high frequency potato cropping soil after seed tuber treatment with antagonistic fluorescent Pseudomonas spp. Phytopathol. Z. 108:207-213.

11. Griffin, D. M., and G. Quail. 1968. Movement of bacteria in moistparticulate systems. Aust. J. Biol. Sci. 21:579-582. 12. Heinrich, D., and D. Hess. 1984. Chemotactic attraction of

Azospirillum lipoferum by wheat roots andcharacterization of someattractants. Can. J. Microbiol. 31:26-31.

13. Hitchcock, P. J., and T. M. Brown. 1983.Morphological heter-ogeneity amongSalmonella lipopolysaccharide chemotypes in silver-stained gels. J. Bacteriol. 154:269-277.

14. Homma, M., H. Fujita, S. Yamaguchi, and T. Iino. 1984. Excretion ofunassembled flagellin by Salmonella typhimurium mutants deficient in hook-associated proteins. J. Bacteriol. 159:1056-1059.

15. King, E. O., M. K. Ward, and D. E. Raney. 1954. Two simple media for the demonstration of pyocyanin and fluorescin. J. Lab. Clin. Med. 44:301-307.

16. Kloepper, J. W., J. Leong, M. Teintze, and M. N. Schroth. 1980. Enhanced plant growth by siderophores produced by plant growth-promoting rhizobacteria. Nature (London) 286:885-886. 17. Kloepper, J. W., J. Leong, M. Teintze, and M. N. Schroth. 1980. Pseudomonas siderophores: a mechanism explaining disease-suppressive soils. Curr. Microbiol. 4:317-320.

18. Lugtenberg, B., R. Peters, P. van der Hoek, and L.vanAlphen. 1975. Electrophoretic resolution of the "major outer membrane protein" ofEscherichia coli K12 into four bands. FEBS Lett. 58:254-258.

19. Mandimba, G., T. Heulin, R. Bally, A. Guckert, and J. Balandreau. 1986. Chemotaxis of free-living nitrogen-fixing bac-teriatowards maize mucilage. Plant Soil 90:129-139.

20. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning, a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

21. Markwell, M. A. K., S. M. Haas, L. L. Bieber, and N. E. Tolbert. 1978. A modification of the Lowry procedure to sim-plify protein determination in membrane and lipoprotein sam-ples. Anal. Biochem. 87:206-210.

22. Marugg, J. D., M. vanSpanje,W. P. M. Hoekstra, B. Schippers, and P. J. Weisbeek. 1985. Isolation and analysis of genes involved in siderophore biosynthesis in plant-growth-stimu-latingPseudomonas putidaWCS358. J. Bacteriol. 164:563-570. 23. Pannekoek, H., I. A. Noordermeer, C. van Sluis, and P. van der Putte. 1978. Expression of the uvrB gene of Escherichia coli: in vitro construction of a pMB9 uvrB plasmid. J. Bacteriol. 133:884-890.

24. Sanderson, K. E., and P. E. Hartman. 1978. Linkage map of Salmonella typhimurium, edition5. Microbiol. Rev. 42:471-519. 25. Scher, F. M., and R. Baker. 1982. Effect of Pseudomonas putida and asynthetic iron chelator on induction of soil suppressive-ness to Fusarium wiltpathogens. Phytopathology 72:1567-1573. 26. Scher, F. M., J. W. Kloepper, and C. A. Singleton. 1985. Chemotaxis of fluorescent Pseudomonas spp. to soybean seed exudates in vitro and in soil. Can. J.Microbiol. 31:570-574. 27. Schippers, B., A. W. Bakker, P. A. H. M. Bakker, P. J.

Weisbeek, and B. Lugtenberg. 1986. Plant growth-inhibiting and stimulating rhizosphere microorganisms, p. 35-49. In V. Jensen, A.Kj0ller, and L. H. S0rensen (ed.), Microbial communities in soil.Elsevier/North-Holland Publishing Co., Amsterdam. 28. Schippers, B., F. P. Geels, 0. Hoekstra, J. G. Lamers,

C. A. A. A.Maenhout, and K. Scholte. 1985. Yield depressions in narrowrotations caused by unknown microbial factors and theirsuppression by selected pseudomonads, p. 127-130. In C. A.Parker, K. J. Moore, P. T. W. Wong, A. D. Rovira, and K. J. Kollmorgen (ed.), Ecology and management of soil-borne plantpathogens. The AmericanPhytopathological Society, St. Paul, Minn.

29. Simon, R., U. Priefer, and A.Puihler. 1983. A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in gram negative bacteria. Biotechnology 1:784-791.

30. Sokal, R. R., and F. J.Rohlf. 1981. Biometry. W. H. Freeman and Co., SanFrancisco.

31. Tsai, C. M., and C. E. Frasch. 1982. A sensitive silver stain for detecting lipopolysaccharides in polyacrylamide gels. Anal. Biochem. 119:115-119.

32. Tsuda, M., and T. Iino. 1983. Transductional analysis of the flagellar genes in Pseudomonas aeruginosa. J. Bacteriol. 153:1018-1026.

33. van der Meide, P. H., M. Dubbeld, and H. Scheliekens. 1985. Monoclonal antibodies to human immune interferon and their use in a sensitive solid-phase ELISA. J. Immunol. Methods 79:293-305.

http://jb.asm.org/